A new study published on 14 April in the journal PLOS ONE has shown that specially trained dogs are capable of detecting positive samples of COVID-19 with 96% accuracy, suggesting the potential of using dogs in large public gathering post-lockdown to actively screen for infection.
The study has demonstrated that even though dogs are, indeed, capable of discerning the virus, it takes careful and extensive training, which is further complicated by the intermingling of different odours in real-life situations.
Since the Working Dog Center at the University of Pennsylvania School of Veterinary Medicine – where the researchers normally work – is currently on lockdown, the team partnered with Pat Nolan, a dog trainer working from a facility in Maryland.
The study enrolled eight Labrador Retrievers and a Belgian Malinois who were not involved in any medical-detection work in the past.
First, the dogs were trained to sniff out a synthetic substance known as the universal detection compound (UDC). Next, the researchers moved on to positive and negative COVID-19 samples that were collected from both adults and children, and deactivated in advance to protect the dogs from infection.
After three weeks of training, the dogs were able to pick up on positive samples with 96% accuracy on average, but were more susceptible to false negatives, which is likely due to the strictness with which the experimental set up categorised what “counts” as accurate.
During the training, researchers faced a variety of challenges, such as the dogs discriminating between individual patients, rather than infection status, and getting thrown off by a sample from a patient who’d tested negative for SARS-CoV-2, but recently recovered from COVID-19.
This prompted the team to conclude that future studies should be conducted by using more diverse samples, and should avoid training dogs repeatedly on the samples from any single individual.
The study is part of another project that endeavours to train dogs to discriminate between individuals with and without COVID-19 disease, as well as those who have been vaccinated, based on volatile organic compounds left on t-shirts.
“That’s something we can carry forward not only in our COVID training but in our cancer work and any other medical detection efforts we do,” said senior author Cynthia Otto. “We want to make sure that we have all the steps in place to ensure quality, reproducibility, validity, and safety for when we operationalise our dogs and have them start screening in community settings.”